Artificial intelligence is changing how organisations recruit, manage and monitor their workforces across the Asia-Pacific region.  

Unlike the European Union’s more comprehensive AI regulatory framework however, most APAC jurisdictions have not introduced AI-specific legislation. Instead, they rely on existing employment, data protection and anti-discrimination laws. Governments have supplemented these frameworks with ethical principles, practical guidance and voluntary governance standards. Although these instruments are not legally binding, employers increasingly regard them as practical compliance benchmarks. 

This approach appears to reflect a preference for flexibility and proportionality. However, the regulatory landscape continues to evolve. Recent developments in Australia and Malaysia, for example, suggest that some governments may be considering a more targeted oversight of AI systems through binding legislation, particularly where those systems may have a significant impact on workers. 

In this article, we explore the regulatory approaches across APAC, the role of non-binding guidance and emerging reforms. Understanding these frameworks is important because they shape how employers assess, govern and deploy AI in the workplace. The insights are based on developments in Australia, Hong Kong, Malaysia, New Zealand, Japan and Singapore.  

These jurisdictions generally regulate workplace AI through existing legal frameworks rather than dedicated AI legislation. 

Employment laws, anti-discrimination rules and data protection obligations continue to apply regardless of whether an employer uses AI-enabled workplace tools. Regulators then supplement these existing legal obligations with guidance and other non-binding instruments. These promote human oversight, bias mitigation, the implementation of governance structures and risk assessments as part of responsible AI deployment. 

Hong Kong’s approach is largely context and risk-based. Existing legal regimes apply to AI-enabled activities, while regulators encourage responsible adoption through non-binding guidance. 

The Personal Data (Privacy) Ordinance (‘PDPO’) is the principal legal framework whenever employers collect, use or process personal data in connection with AI systems. Anti-discrimination legislation also applies, together with sector-specific requirements. 

In parallel to this existing framework, the Office of the Privacy Commissioner for Personal Data (‘PCPD’) has issued several influential, non-binding guidance documents for organisations that are relevant to workplace AI. This includes: 

• The PCPD’s ‘Checklist on Guidelines for the Use of Generative AI by Employees’. This sets expectations for internal governance, privacy compliance, human oversight and bias mitigation when employees use generative AI tools.  

• The PCPD’s ‘Artificial Intelligence: Model Personal Data Protection Framework’. This recommends that organisations adopt an AI strategy and governance structure (for example, an AI governance committee), conduct comprehensive risk assessments (including privacy impact assessments) and implement measures to ensure ongoing PDPO compliance when preparing, training or deploying AI systems. 

The Hong Kong Government has also issued the non-binding ‘Ethical AI Framework’. This provides practical guidance on responsible AI governance, assessment methodologies and ethical principles. 

Singapore has adopted a sectoral, risk-based and outcome-oriented approach. It is anchored in existing laws, while being supplemented by non-binding frameworks and regulator guidance. In fact, non-binding measures are very much central to Singapore’s AI governance model.  

The Personal Data Protection Act 2012 (‘PDPA’) establishes core legal obligations for organisations that use AI systems involving personal data. Employers must comply with requirements relating to: 

• notification and consent; 

• purpose limitation; 

• data accuracy; 

• data protection; 

• retention; 

• access and correction rights; 

• data breach notification; and 

• organisational accountability. 

Employers deploying AI systems, such as algorithmic screening, productivity monitoring tools, or biometrics, must comply with these PDPA obligations and demonstrate appropriate governance, including risk assessments, policies, training, and vendor management. 

Beyond the PDPA, employment-related protections also arise through the ‘Tripartite Guidelines on Fair Employment Practices and the Fair Consideration Framework’. These frameworks require fair, merit-based employment decisions regardless of whether an employer uses AI tools for these. 

Singapore is also progressing Workplace Fairness Legislation, which is expected to introduce statutory protections against workplace discrimination. AI enabled practices in recruitment and employment will need to comply with these statutory standards once in force. The legislation is expected to take effect by the end of 2027. 

Alongside these legal obligations, Singapore has developed a particularly mature governance ecosystem. Key guidance includes: 

• the ‘Model AI Governance Framework for Agentic AI’; and 

• the ‘Implementation and Self-Assessment Guide for Organisations’. 

Together, these instruments promote transparency, explainability, accountability, robustness and human oversight. Sector regulators have also published complementary guidance for AI uses within their industries. 

New Zealand has adopted a light-touch and principles-based approach. Rather than introducing dedicated AI legislation, the Government continues to rely on existing employment, human rights and privacy laws, voluntary frameworks and regulatory oversight. 

Examples include the following: 

• The ‘Algorithm Charter for Aotearoa New Zealand’, which commits signatories to principles of transparency and fairness, and ensures that New Zealanders can have confidence in how government agencies use algorithms. 

• The ‘Public Service AI Framework’, which provides guidance to support the responsible development and deployment of AI across public sector agencies. 

These initiatives focus on transparency, fairness and public trust in the use of algorithms and AI systems. 

Like the above jurisdictions, Japan has no comprehensive regulation specific to the use or deployment of AI in the workplace. Instead, employers must comply with existing employment, anti-discrimination and data privacy laws when using AI tools. 

Japan’s dedicated AI statute – the Act on the Promotion of Research and Development and Utilisation of AI-Related Technologies – has not fundamentally changed this position. In force since September 2025, this legislation sets out a broad AI governance framework. It is designed to advance AI adoption and establish basic principles, including transparency and the protection of citizens’ rights. It does not, however, impose any directly enforceable obligations on employers, nor any penalties.  

Whether to deploy AI-driven HR technology is therefore left to the employer’s discretion in Japan, so long as it is done within the limits of existing legal frameworks. 

In the workplace, AI adoption appears most active in recruitment and performance evaluation. In recruitment, AI interview systems assess behavioural characteristics from recorded responses. In performance evaluation, AI can support a more quantitative assessment of capability. For these common use cases, the relevant obligations arise through the following existing regulations: 

• The Equal Employment Opportunity Act, the Act on Comprehensive Promotion of Labour Policies and the Act to Facilitate the Employment of Persons with Disabilities. These prohibit sex discrimination, age limits and disability discrimination in recruitment and hiring. 

• The Act on the Protection of Personal Information. This governs specification of the purpose of use, notice, use restrictions and special-care-required personal information. The legislation is being amended to regulate facial feature data separately. 

• The Employment Security Act and its related guidelines. These limit the use of job-seeker personal information to what the business purpose requires. The regulations also provide that recruiting organisations should not collect information on race, family origin, beliefs or union membership. The guidelines carry no penalty, but are enforced through administrative guidance. 

• In the context of performance evaluation, the Labour Standards Act and the Labour Union Act. These prohibit treatment based on nationality, creed, social status, union membership or union activities. 

The use of AI does not alter these legal standards. A hiring or evaluation decision may be unlawful where the tool embeds bias linked to a protected attribute. It may also be unlawful if the underlying data, or the manner of its collection, breaches the requirements above. Furthermore, an evaluation that amounts to an ‘abuse of discretion’ may render the measure void and give rise to damages.  

Unlike some other jurisdictions, Japan has no equivalent right for employees not to be subject to solely automated decision-making. There is therefore no legal requirement for employers to carry out human review of AI outputs, underlying datasets or AI-generated interview questions. Nevertheless, such review remains important to mitigate potential liabilities under the regulations mentioned above. 

Malaysia also relies to a large extent on non-binding instruments, although recent developments paint an evolving picture.  

The Government has issued the ‘National Guidelines on AI Governance and Ethics’ which are intended as voluntary guidance for industry players whilst the Government develops laws to regulate the use of AI. The guidelines recommend that employers: 

• notify employees when AI is used in the workplace; 

• respect employee privacy rights, as required by law; 

• align AI deployment in the workplace with HR policies; and 

• ensure bias-free AI. 

Building on this, the Personal Data Protection Commissioner in Malaysia is also developing specific guidelines for the use of automated decision-making in processing personal data. Once these guidelines have been developed, they are expected to further influence how employers can use AI in the workplace.  

Despite the current reliance on non-binding frameworks, however, recent developments in Malaysia suggest that a more formal framework may be emerging. In particular, on 10 July 2026, the Malaysian National AI Office published a public consultation paper on a proposed AI Governance Bill. The proposal adopts a principle-based and risk-based approach and would establish a Central AI Authority to oversee AI governance.  

If enacted, the Bill would introduce stricter requirements for AI systems deployed, used, or developed in Malaysia, including those deployed by an organisation for employment-related purposes. High-impact AI systems could be subject to more stringent requirements, although the consultation paper is silent on whether this expressly includes employment-related AI tools.  

The proposed framework places particular emphasis on: 

• human agency and human rights; 

• transparency and explainability; 

• accountability and effective redress; 

• the safe and secure use of AI systems; and 

• responsible data governance.  

Although the Bill is still in the early legislative stages, it signals a potential shift towards formal AI regulation in Malaysia. Employers should begin to assess AI tools currently utilised in employment settings in preparation for potentially stricter compliance requirements. 

Finally, and alongside these legislative efforts, the Government has also been strengthening its infrastructure for AI oversight. Malaysia’s Ministry of Digital formally established the National AI Office (‘NAIO’) through the launch of the national AI lead agency, AI Malaysia Berhad (‘AIMB)’. AIMB is poised to carry out the following six strategic functions: 

• strategy, policy and foresight; 

• national AI implementation; 

• trusted AI governance; 

• partnerships and international cooperation; 

• AI capability development; and 

• the Malaysia AI Safety Institute (‘MY-AISafe’). 

Together with the unveiling of the ‘National AI Action Plan 2026-2030‘, these developments demonstrate Malaysia’s continued prioritisation of building a sustainable AI ecosystem and strengthening its position as a regional digital technology hub. 

Australia broadly follows the regional model outlined above. While it has not adopted a comprehensive workplace AI regulatory framework, many aspects of AI deployment are already regulated through existing employment and workplace laws.  

Relevant obligations arise through: 

• unfair dismissal (including redundancy) laws; 

• anti-discrimination legislation; 

• adverse action protections; 

• work health and safety requirements;  

• consultation obligations under modern awards and enterprise agreements; 

• protections for employee-like ‘gig’ workers in relation to unfair deactivations; and 

• privacy and workplace surveillance laws. 

With that said, more targeted regulation is beginning to emerge. Notably, New South Wales has introduced legislation confirming that existing work health and safety duties apply to risks arising from digital work systems (broadly defined as an algorithm, artificial intelligence, automation or online platform) in use by the business, including algorithmic management and AI-enabled allocation of work. 

Whilst the Australian Government has not yet introduced a dedicated AI Act, it has introduced a National AI Plan and also established the Office of AI in July 2026, to accelerate the implementation of a new set of Australian Standards for AI. While the initial focus is on large data centres and AI training rather than workplace specific issues, these developments signal a broader interest in AI governance and regulation at the federal level.  

Like other jurisdictions in the region, the Australian Government has also issued non-binding guidance for organisations deploying AI technologies. This includes: 

• ‘Guidance for AI Adoption: Foundations‘ which is intended for organisations in the early stages of using AI or who are deploying ‘low risk’ AI systems.  

• ‘Guidance for AI adoption: Implementation Guidance‘ which is aimed at organisations deploying ‘complex high-risk’ AI systems.  

It has also established a tripartite AI Employment and Workplaces Forum to facilitate discussion between representative employers, unions and governments regarding workforce challenges arising from AI implementation. 

A gap analysis of Australia’s workplace law legislation is currently being undertaken by the Australian government, which may lead to future legislative reform in the area.  

At the Federal level, trade unions, led by the Australian Council of Trade Unions (‘ACTU’), have advocated for mandatory ‘AI Implementation Agreements’ that would require employers to consult with staff before introducing new AI technologies and address issues such as job security, skills development, retraining, and transparency over technology use, amongst others.   

Recent bargaining trends indicate that trade unions are increasingly focussed on securing AI-related clauses within enterprise agreements. The number of agreements including AI-related clauses has increased throughout 2026, with common themes addressing issues such as consultation, transparency, protection against harmful surveillance and job displacement, access to skills and training, and stronger enterprise-agreement protections.  

Australian courts and tribunals have also now taken steps to respond to the growing use of generative AI in litigation, including most recently, the Federal Court of Australia and the Fair Work Commission. Although approaches differ between jurisdictions, there is a growing expectation that parties remain accountable for AI-assisted work product and take reasonable steps to verify its accuracy and appropriateness. 

While Australia’s approach remains less prescriptive than some overseas jurisdictions, the direction of travel is clear: employers adopting AI should expect increasing scrutiny of governance, consultation, transparency, safety and risk-management practices. 

Employers in APAC cannot assume that the absence of AI-specific legislation means workplace AI is unregulated. Existing employment, data privacy and anti-discrimination laws already apply to AI-enabled processes, and organisations must comply with local frameworks. 

At the same time, regulators across the region are increasingly supplementing these legal frameworks with non-binding guidance, ethical principles and governance expectations.  

Against this backdrop, several practical measures are emerging as important components of responsible AI adoption, both from a compliance and workforce relations angle. We explore these here, setting out a checklist of these measures for APAC employers, covering governance, risk assessment, privacy, training and ongoing oversight.